Holz, List, Zentner, Cardoza, and Zentner (2020) on Nudging Tax Compliance

Justin E. Holz, John A. List, Alejandro Zentner, Marvin Cardoza, and Joaquin Zentner, “The $100 Million Nudge: Increasing Tax Compliance of Businesses and the Self-Employed Using a Natural Field Experiment.” NBER Working Paper 27666, August 2020 (pdf here). 

• The tax authorities send you a message… 

• ...maybe the message just happens to mention the upcoming tax filing deadline: the control arm 

• ...maybe the message also notes the potential for your tax evasion to be publicized – one treatment arm, designed to increase the salience of social penalties for being a tax cheat. 

• ...maybe, instead of highlighting the public nature of identified tax evasion, the message notes the potential for your tax evasion to result in imprisonment – another treatment arm, designed to increase the salience of criminal penalties attached to tax evasion. 

• And for each of the three arms noted above, we can take half of the letter recipients and also mention in the letter that the tax authorities are prepared to potentially view mistakes in tax declarations, even honest mistakes, as intentional. (The notion is to frame evasion as a sin of commission, not of omission.) Mistakes imply that you were trying to be a tax cheat! 

• This very natural field experiment is conducted in the Dominican Republic, circa 2019, n≈56,000 firms and n≈28,000 self-employed people; tax evasion reportedly is rife in the Dominican Republic. 

• The field experiment applies to business entities subject to the corporate income tax and to self-employed taxpayers subject to the individual income tax. 

• The various messages are sent shortly before the tax filing deadline. 

• The threat of public disclosure of tax evasion dissuades tax evasion for both firms and individuals. 

• The “prison” message also reduces evasion, and for firms, about twice as effectively as the “publicity” message. 

• Framing evasion as an active choice, a sin of commission, in itself (without publicity or punishment prompts), does nothing (or worse than nothing), and likewise is ineffective if it is paired with the publicity notice. 

• But the combination of “intentional” framing with the prison message doubles the impact of the prison message. 

• The effectiveness of the interventions seems to arise from a decrease (by 20%) in potential taxpayers who declare (falsely, presumably) that their income is below the minimum required for taxation. 

• Large firms drive the reduced tax evasion – there is little or no compliance gain from the smallest 60% of taxpayers.

Špecián (2019) on True Preferences

Petr Špecián, “The Precarious Case of the True Preferences.” Society 56(3): 267-272, June 2019.

• Are voluntary trades (with no spillovers) Pareto improving? To assert that they are suggests a belief in coherent preferences. This is the standard economics approach: choices reveal underlying, true preferences.

• Behavioral evidence undermines the belief that choices reveal true preferences. Nonetheless, nudges aim to make people better off “as judged by themselves," so there is an appeal to some underlying true preferences from which those judgments emerge.

• But when can we know that people's un-nudged choices do not reflect their true preferences? How can we determine which preferences should be disregarded or questioned? Signposts of potential irrationality might include rare choices, intertemporal choices, choices involving temptation goods…

• Can we rely upon cold-state (as opposed to hot-state) preferences, or long-run, considered preferences, as somehow being "true" preferences? But maybe these proclaimed "more considered" preferences reflect social norms and are just marketed for public consumption, even as people really want to be Mr. Hyde, not Dr. Jekyll. 

• Perhaps Sugden (2008) is right, perhaps there are no underlying coherent preferences at all.

• Nudgers might believe that their interventions make matters better in practice even if these fundamental questions about how to judge individual welfare remain unresolved. [That's pretty much my view, too -- JL] Nudgers should be humble, and wary of overconfidence.

Duckworth, Milkman, and Laibson (2019) on Willpower and Beyond

Angela L. Duckworth, Katherine L. Milkman, and David Laibson, “Beyond Willpower: Strategies for Reducing Failures of Self-Control.” Psychological Science in the Public Interest 19(3): 102-129, 2019.

• Higher self-control is associated with many good things -- but people struggle with self-control.

• Various models (such as dynamically inconsistent preferences) have been proposed to try to capture the workings (and failures) of self-control, and many strategies are deployed to try to bolster self-control. Exhortations to show more willpower, to just say no, are not that effective. That is, relying on willpower alone, as opposed to other strategies, is almost to give the game away.

• Strategies that try to make willpower more effective can usefully be categorized. In particular, some try to alter the situation (Situational strategies), and some try to target how the decision maker thinks about a given situation (Cognitive strategies). Further, some are undertaken by the decision maker herself (Self-Deployed strategies) and others are enacted by various outside actors, choice architects, including employers and regulators (Other-Deployed strategies). Self-Deployed strategies require that the decision maker understand, at least to some extent, her self-control shortcomings -- in the usual terminology, she must be somewhat sophisticated (not naive) about her intertemporal inconsistencies.

• Four categories of self-control strategies emerge within this two-dimensional framework, then: Self-Deployed Situational Interventions; Self-Deployed Cognitive Interventions; Other-Deployed Situational Interventions; and, Other-Deployed Cognitive Interventions. The authors list and discuss various strategies within each of these categories.

• Self-Deployed Situational Interventions attempt to manage the environment to reduce the need for self-control; such strategies include: deleting a game from your phone or employing other commitment devices (or commitment contracts) that reduce temptations or raise their effective prices; temptation bundling; removing temptations from view (which is one way to raise their effective price); and behavior therapy, such as removing "cues" from your environment when fighting an addiction.

• Self-Deployed Cognitive Interventions include: goal setting (which might be undertaken in public or in private); dividing the goal into bite-sized subgoals, and allowing for small wins along the way; explicit planning and implementation intentions (developing a detailed map of how a goal will be achieved); intensive self-monitoring (such as weighing yourself every day); psychological distancing from temptations or from interpersonal problems; mindfulness and cognitive therapy (generally provided in the form of Cognitive Behavioral Therapy).

• Other-Deployed Cognitive Interventions (which might not be recognized by the targets) include: descriptive social norms (like energy use by others in your neighborhood); social labeling (which might motivate you to live up to your identity, as a green consumer, say); age-processed renderings or other techniques to make one's own future self more relatable; and fresh-start framing, as with new year's resolutions.

• Other-Deployed Situational Interventions include: hard paternalism (e.g., social security, high sin taxes); microenvironments (healthy food placed at eye level); default settings or active choice mandates; requirements for choices to be made in advance; and planned interruptions (such as a mandated break on a slot machine, say).[Incidentally, as Loewenstein points out, hard paternalism does not bolster self-control but eliminates the need for self-control.]

Damgaard and Nielsen (2018) on Nudgication

Mette Trier Damgaard and Helena Skyt Nielsen, “Nudging in Education.” IZA DP No. 11454, April 2018.

• Education decisions often involve current costs and greatly delayed benefits. Some decisions are made by “agents” – parents and teachers – on behalf of their “principals,” the students. 

• Many education decisions – such as whether to drop out of school or go to college – are not regular, everyday situations where feedback allows people to adapt to something tolerably optimal over time. 

• Can we be confident that education decisions are well-made? Evidence suggests high returns to more education. The list of potential behavioral biases influencing education decisions is long: self-control shortcomings, limited attention, loss aversion, default effects, social norms, underconfidence… Recall John Stuart Mill: "Those who most need to be made wiser and better, usually desire it least, and if they desired it, would be incapable of finding the way to it by their own lights."

• Perhaps people can be nudged in ways that will improve their choices – and improve them from their own point of view. 

• The targets of nudges can be young scholars, their parents, or their teachers. 

• The authors find 122 studies of field implementations of nudges in education. 

• Opt-in v. opt-out text messages to parents: only 7.8% opt-in, but only 3.5% opt out. Further, outcomes (grades, staying in school) improve with the opt-out default. 

• Framing financial aid as a tuition waiver versus a loan: the waivers steer a lot more law school grads to public interest work. Monetary transfers to families can be framed as being for education, and administered through schools. 

• Mixed results have been reported from framing grades as losses (as in Shrader, Wooten, White, et al., “Improving Student Performance through Loss Aversion”).

• Trying to take advantage of peer group effects is tricky, and sometimes counterproductive. Social comparison nudges can be both informational and motivational – and they can backfire.

• Intermediate assignments and deadlines seem to often raise grades. 

• Self-imposed specific task-based goals also seem helpful, along with reminders of the goal. Unrealistically high goals are demotivating. 

• Reminders to parents about their children’s education often seem helpful, too – likewise with providing informative updates to parents.

• Information provision about the returns to schooling seems to have some purchase in developing countries.

• “Boosting” skills like grit and goal-setting holds some hope for improved educational outcomes.

• Prizes: “The public can encourage the acquisition of those most essential parts of education by giving small premiums, and little badges of distinction, to the children of the common people who excel in them.” (Adam Smith) But the older the children, the less effective the prize nudge, and (like social comparisons), prizes can crowd out internal motivations. 

• Growth mindset and social belonging nudges seem to work OK…

• Nudging will only overcome a binding constraint for a (small?) subset of students. 

•  In general, the long-term effects of nudgication remain unknown.

Golman, Loewenstein, Moene, and Zarri (2016) on Belief Consonance

Russell Golman, George Loewenstein, Karl Ove Moene, and Luca Zarri, “The Preference for Belief Consonance.” Journal of Economic Perspectives 30(3): 165-188, Summer 2016 [pdf].

• People like to have beliefs that accord with the beliefs of others. Sharing beliefs enhances our connection to our group. 

• Much of world conflict is about beliefs – often about rather subtle differences in beliefs. Recall that people protect beliefs in which they have invested heavily. 

• Conflict over small differences in beliefs might arise because our beliefs are most threatened by those who are otherwise similar to us. 

• People do not like to have their beliefs challenged, so media have incentives not to challenge beliefs.

• Beliefs might come first, and only then do we develop the “rational” reasons that we hold them; see Jonathan Haidt, The Righteous Mind: Why Good People Are Divided by Politics and Religion, Pantheon, 2012. 

• Belief consonance can be self-reinforcing: when someone stubbornly refuses to agree with me, I can attribute her stubbornness to her own interest in protecting her initial beliefs – and therefore I do not have to reconsider my own beliefs. 

• In the interest of possessing consonant beliefs, all people might believe X, but believe that everyone else believes “not X”: “pluralistic ignorance”. See Timur Kuran, Private Truths, Public Lies: The Social Consequences of Preference Falsification, Harvard, 1998. 

• In trust and dictator games, people are more generous when paired with members of their own political party.

• Two (complementary?) approaches to the enticements  of belief consonance: (1) desire to match beliefs with a group that you are in, or want to join; and (2) desire to maintain desirable beliefs about yourself. 

• On the whole, belief consonance probably is detrimental to society.

Horton and Zeckhauser (2016) on Peer Effects in Production

John J. Horton and Richard J. Zeckhauser, “The Causes of Peer Effects in Production: Evidence from a Series of Field Experiments.” NBER Working Paper No. 22386, July 2016 [ungated pdf version here].

• A worker’s productivity is influenced by the productivity of her co-workers. Why? 

• Do low-productivity workers fear punishment, because their slackerdom imposes more work on others, or because others simply view the provision of low effort as unfair? Do workers have a preference to not be unproductive? Does the performance of other employees signal to a worker the employer’s expectations?

• The field experiments involve hiring workers online (MTurk) to label images; the workers do not know that they are taking part in an experiment. 

• Among the findings are that peers will punish slackers even when the slackers do not impose any harm on the other workers; and, that equity concerns motivate such punishment, through a suspicion of low effort from slackers. Workers therefore want either to avoid being slackers, or, perhaps, to avoid being perceived as slackers. 

• Workers (the subjects in the experiments) label images and evaluate the labelling performance of other workers. Evaluations involve a recommendation as to whether the peer worker should be paid, and also, a suggestion of how a 9-cent bonus should be split between the evaluator and the evaluated. The recommended bonus splits are implemented. A recommendation that a worker not be paid, though not implemented, is taken to be a type of punishment of that worker.

• All of the subjects also evaluate the work of either a (specific) high effort or a low effort worker, based on work from a previous experiment. A description of the four treatments in the field experiment follows:

Baseline: Workers are shown either a very good labelling job, or a minimal effort version. They then choose to take the task (if indeed they want to), and start labelling. 

Punish: After seeing a sample of excellent work, this treatment proceeds like Baseline, except that then some good or bad work requires peer evaluation. 

Peers: After evaluation, subjects then are given a second labelling task. Will evaluating a good or a bad job (done by someone else) alter the worker’s second-round performance? 

Explicit: This is like Peers, except that workers are told to produce only two labels per image. But when they evaluate others, some of the subjects see an image with the requisite two labels, while the others see an image with 11 labels (produced by some Stakhanovite). Note that this excessive production explicitly contravenes the injunction to provide only two labels.

• And the results...

Baseline: If you are shown a high-quality sample, you are less likely to take the job, but if you take it, you increase your effort. 

Punish: Peer evaluation leads to lessened punishment for good as opposed to bad work; workers who themselves produce good work punish more. 

Peers: If you evaluate a strong worker, your subsequent work is stronger, and the effect is more pronounced if you yourself are a  high productivity worker. 

Explicit: Workers shown against-the-rules overproduction do not punish it, and many workers themselves switch to excessive output.

Kremer and Levy (2008) on Peer Effects and Alcohol

Michael Kremer and Dan Levy, “Peer Effects and Alcohol Use Among College Students.” Journal of Economics Perspectives 22(3): 189-206, Summer, 2008.

• The data come from a large state university that randomly assigns some roommates. The finding: males assigned roommates who drank in high school had a lower GPA. 

• By using high school characteristics, we can rule out that the GPA connection stems from common shocks to the roommates. 

• For the most part, studies don’t find much support for the idea that academic characteristics affect the academic performance of roommates. Kremer and Levy likewise find no effect of academic background or social background on roommate academic achievement. 

• The GPA decline for males is about .27 if the roommate drank in high school. The drop is especially bad for weak students, and for those who themselves drank in high school, and it’s much greater in the second year (though you are no longer roommates). 

• Roommates who are not randomly assigned, but who choose each other, do not see falls in GPA from roommate drinking history. 

• Reducing the drinking of one student can reduce the drinking of others. But collecting non-drinkers in substance-free housing will concentrate the drinking students, with a negative effect on overall GPAs.

Levitt and List (2007) on Laboratory Experiments

Steven D. Levitt and John A. List, “What Do Laboratory Experiments Measuring Social Preferences Reveal about the Real World?” Journal of Economic Perspectives 21(2): 153-174, Spring, 2007.

• Five laboratory games that reveal something about pro-social behavior: (1) the Ultimatum Game; (2) the Dictator Game; (3) the Trust Game; (4) the Gift Exchange Game (where the first player requests a level of “effort” from the second player -- pdf here); and (5) the Public Goods Game (good video here). In the real world, it can be hard to differentiate between pro-social preferences and sophisticated self-interest seeking. 

• Model notation: action choice a; wealth W; stakes (or value) v; moral cost M; social norms n; and scrutiny s. 

• The higher the negative financial externality an action imposes on others, the higher the moral cost M is taken to be. (Levitt and List also assume that this externality increases with the stakes v). M is higher the greater the deviation between action a and the social norm n. M also is raised by increased scrutiny s. 

• Individual utility is U(a, v, n, s) = M(a, v, n, s) + W(a, v). Higher stakes v can raise W while raising M, too, but the authors assume that W rises more quickly with v. (Another interpretation might be that the norm changes with v, so that selfish behavior receives more social imprimatur.) 

• Scrutiny is different and typically more intense in the lab, exaggerating pro-social behaviors. (Alternatively, scrutiny from one’s children or other family members has no lab parallel.) Further, lab participants might believe that an experiment demands some pro-social behavior. 

• Behavior might be sensitive to factors that unavoidably vary between the lab and the real world: the experimenter cannot fully control the context. Participants bring context with them, and hence are playing a different game. 

• Lab participants self-select, directly or indirectly, while market participants self-select, too.